The rapid development of the ChatGPT voice robot is demonstrating the potential of artificial intelligence in the financial sector. As research by finews.com shows, there is a window of opportunity for small providers in particular to make big leaps.
Even Sergio Ermotti is talking about the A-word these days. At an industry meeting last week, the UBS CEO, who is fully absorbed in the integration of Credit Suisse, spoke about the potential of artificial intelligence (AI).
The head of the soon-to-be $5 trillion wealth manager explained in broad terms how the technology can be used to deliver a better quality of service to clients.
Reto Ringger, CEO of the sustainability-focused Swiss asset management bank Globalance, is more explicit. «AI will be disruptive for our industry,» he said in an interview with finews.ch. ChatGPT plug-ins are already capable of setting up individual investment portfolios within seconds. «This technology will improve rapidly. Who will still want to pay their bank 100 basis points for the same service?» he muses.
Globana is Just the Beginning
But Ringger doesn't want to wait for customers to find their answers. He says that within the bank every division of the company has to embrace the new technology's potential.
Globalance ventured ahead with an initial application with the launch of its «AI employee Globana,» an avatar providing information on the potential of the technology from an investor's perspective in short videos. Ringger freely admits it's more of a gimmick. But he was surprised at how quickly and cost-effectively Globana went live and is already thinking about expanding it. The future AI employee could inform customers according to their interests, the CEO hopes.
In general, he sees smaller players as having an advantage in AI applications. The implementation speed is much higher, and costs for these technologies falling simultaneously, helping to keep up with the rapid development.
Two Camps
There are currently different camps in Swiss finance in the view of Sandro Schmid and Nourdine Abderrahmane, partners at technology consulting firm LPA. «There is a faction that's very positive about the new possibilities: people see opportunities to reduce costs and improve the user experience via AI,» says Abderrahmane. Generative AI is also a wonderful tool to better manage the wealth of investment data, he adds.
Still, skeptics emphasize the relationship in private banking is based on personal client relationships.
«We believe the two viewpoints aren't mutually exclusive. AI is ideally suited to assist the human advisor. From an institution's point of view, such a tool should not be prematurely closed off,» says Schmid. This is all the more important because the topic is also gaining traction among authorities, including the Swiss Financial Market Supervisory Authority (Finma). «They are interested in banks openly addressing the issue. It would be fatal for Switzerland not to take advantage of the opportunities,» Schmid said.
Lack of Swiss Data
Selma Finance, one of the first digital asset managers in the Swiss financial industry, can't be accused of this. In June, the fintech raised around 1.3 million Swiss francs ($1.4 million) via crowdfunding to make digital advice fit for the ChatGPT era.
Selma CEO Patrik Schaer explains the company is currently testing use cases for the technology. One example is using AI to determine how to combine Pillar 3a private pensions with other investment goals.
Even so, there are hurdles to overcome. There is a lack of accessible data on pension provisions in Switzerland from which a machine could learn, and building an AI application from scratch without existing plug-ins is costly, Schaer explains.
No More One-Size-Fits-All
The first applications of AI in banking will not be in asset management, but in compliance, in the view of Schmid and Abderrahmane. Here, the corresponding tools are already in use in the fight against money laundering and KYC applications. Risk management is also an important playground for technology, along with process automation.
They say better processes also benefit the customer and algorithms could help make more targeted approaches, increasing the bond between the customer and the institution. With that, banks can say goodbye to the «one size fits all» process.
For investments, Abderrahmane and Schmid see great opportunities in portfolio optimization. «The models are getting better and better, and at the same time, a much larger amount of data can simply be made accessible and processed in a way that makes sense for the advisor.»
Nightly Risk Check
Such approaches have quietly been in use at UBS for years. «UBS Advice» is one tool used in the mandate area that checks investment portfolio risk every night and then makes suggestions to clients.
The «Next Best Action» approach, in operation at UBS Switzerland since 2021, is more recent. It analyzes client behavior, translating it into recommendations, which the bank advisors can take to their clients. According to reports, it's a relatively simple model that only makes marginal use of machine learning.
Focus on Efficiency
It's no coincidence UBS is categorizing its efforts in this area under the heading «AI and automation.» If it's investing in smart machines at all, it's to make processes more efficient for customers and employees while making large volumes of data usable.
Considering the combined UBS will have to save around $8 billion by 2027, with expected integration costs of up to $10 billion, it becomes obvious the scope for AI experiments is likely to be rather limited.
